A use of uninitialized value vulnerability in Tensorflow
tensorflow
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
tensorflow
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
|
|
| tensorflow-gpu PyPI |
|
|
TensorFlow is an open source platform for machine learning. In affected versions TensorFlow's Grappler optimizer has a use of unitialized variable. If the `train_nodes` vector (obtained from the saved model that gets optimized) does not contain a `Dequeue` node, then `dequeue_node` is left unitialized. The fix will be included in TensorFlow 2.7.0. We will also cherrypick this commit on TensorFlow 2.6.1, TensorFlow 2.5.2, and TensorFlow 2.4.4, as these are also affected and still in supported range.
From the CVE record · GitHub_M
TensorFlow’s Grappler optimizer can use an uninitialized variable when a saved model’s `train_nodes` vector contains no `Dequeue` node. An attacker who can supply a saved model for optimization could trigger this condition. The input does not specify network access, authentication requirements, or user interaction.
Check deployments using the PyPI packages tensorflow, tensorflow-cpu, or tensorflow-gpu in versions before 2.4.4, 2.5.0–before 2.5.2, 2.6.0–before 2.6.1, or 2.7.0 release candidates.